digital_audiofandomcom-20200213-history
Bit Rate
Bit rate (not to be confused with bit depth) is a measure of the number of bits (binary digits) transmitted/read/written/reproduced or encoded in a given length of time. This is essentially an equivalent factor to the detail of a given data file. Bit rate is usually measured in kilobits per second (kbps or kb/s) In audio terms, bit rate is a result of combining the sample rate of a given piece of audio and its bit depth. If we know the speed at which samples are played back (E.G. 44,100Hz), we know the bit depth (bits per sample), and we know the number of channels (usually 2, for stereo sound), then we can calculate the bit rate: *Bit Rate = Bit depth * Sample Rate * #channels To calculate a predicted filesize, we can multiply the bitrate by the duration of the file, dividing by 8 to convert the units to bytes: *Resultant File Size = ((Bit depth * Sample Rate * #channels) * duration ) /8 Bitrate as an indicator of Audio Quality Much more important to actual data fidelity than the format or filetype of the audio (e.g. mp3 or wav), the bitrate is actually a quite clear indicator of the resultant quality. In C. Brown Listening Tests, 2014, participants listened to a series of audio files - some identical - comparing CD quality to various lossy bitrates, and some bitrates to one other. The threshold beyond which the majority of participants could actually tell the difference was 112kbps mp3. Across the board once the bitrate had climbed beyond 128kbps, even in a critical listening environment and especially at home, members of the public began to struggle to pick out any difference at all. Industry professionals had an unsurprisingly higher threshold, but whether this is due to training or an innate ability is hard to say. Most likely it is a combination of experience and available listening technology. Therefore can it be assumed that a higher bitrate audio file will always sound better than the same audio at a lower rate? Not necessarily. For example, the AAC codec provides a slightly higher quality at similar bitrates, so it would follow that the threshold could go lower than 112 in theory, although the study would indicate that is not the case because the margin between 112 and the next codec down (98kbps) seems to be substantial enough to produce an audible drop in quality. Additionally, a previously compressed audio file can be "upsampled", even being re-encoded to FLAC, but this will not restore lost information. Fig. 1 visually illustrates this problem. If an image is shrunk to a very small size, saved, then re-opened and expanded back to its original size, some of the pixels which were originally in place may have moved, been removed, or been otherwise compressed. Although the file will get bigger again (in physical and data size) and (back in the audio realm) have a higher bitrate, the quality will remain at its lowest point. This creates a very difficult situation for the consumer, and unfortunately there are seemingly few ways around this problem. One long-term solution however is simply attempting to further public awareness of the issue and the implications of lossy data compression with hopes that this will create an upward trend in average quality. Bitrate Summary *Bitrate is the measure of the number of bits in a data stream, usually measured in bits or kilobits per second (bps or kbps) *Calculated as a combination of bit depth (binary word length), sample rate, and the number of audio channels (two for stereo, one for mono) *Bitrate can be used as a rough guide for audio quality, as lower bitrates will certainly be of lower quality. However it is up to the consumer to judge each file for themselves and find their comfortable threshold